Novel duck reovirus exhibits pathogenicity to specific pathogen-free chickens by the subcutaneous route

To study the pathogenicity of new duck reovirus (NDRV) to chickens, eighty 3-day-old SPF chickens were equally divided into two groups. The experimental group was inoculated with a NDRV challenge strain of 100 μL (10–5.00 ELD50/0.1 mL) by the subcutaneous (s.c.) route, and the control group was inoculated with 100 μL of sterile phosphate-buffered saline (PBS) by the same route. In the experimental group, chickens exhibited introflexion of claws, performing of splits, stunting syndrome, weight loss and death. Gross lesions such as enlargement and yellowish-white focal necroses were observed in the liver and spleen. Microscopic changes were typical including varying degrees of hepatocyte steatosis and necrosis, splenic lymphocyte necrosis, interstitial pneumonia. Viral loads were detected in lung, liver, heart, spleen, duodenum, burse and kidney. The liver and spleen viral loads remained a much higher level and maintained for a longer time, suggesting that these tissues might be the target organs. In summary, NDRV can cause systemic infections and death in chickens, which indicated that chickens may be infected by NDRV in poultry production.

. In detail, all infected chickens at 3 and 6 dpi, as well as 4/5 chickens at 9 dpi showed severe lesions in the liver and spleen. Hepatomegaly and brittleness were seen in the liver, and there were many yellowish-white focal necroses of variable size, on the surface or in the parenchyma of the liver (Fig. 1C). The spleen was swollen, showed haemorrhages, and multifocal yellowish-white necroses, up to 3 mm in diameter, present on the surface or in the parenchyma (Fig. 1D). No gross lesions were found in other organs of experimental group or in any organ of the control group. Table 2. At 3 dpi, four out five (4/5) chickens showed mild interstitial pneumonia with inflammatory cell infiltration and haemorrhage (Fig. 4A). Moderate interstitial pneumonia was observed in the lungs of 4/5 infected chickens at 6 dpi ( Fig. 4B). Mild congestion and lymphocyte infiltration were observed in 4/5 chickens at 9 dpi (Fig. 4C). In the control group, the detection of erythrocytes in parabronchi and air capillaries was attributed to the killing and sampling procedure. (Fig. 4D). Livers of all infected chickens at 3 dpi showed severe hepatocyte steatosis and necrosis ( Fig. 4E). At 6 dpi hepatocyte degeneration, necrosis and inflammatory cell infiltration were more severe (Fig. 4F). Focal bleeding was obviously observed (Fig. 4F). At 9 dpi, inflammatory cell infiltration was mild (Fig. 4G). No lesions were found in livers of the control group (Fig. 4H). Hearts of 4/5 infected chickens showed moderate microscopic lesions at 3 dpi and lesions began to improve at 6 dpi and 9 dpi. At 3 dpi, there was moderate inflammation and cardiomyocytes showed granularity of the cytoplasm (Fig. 4I). Mild inflammatory cell infiltration was found in 4/5 infected animals at 6 dpi ( Fig. 4J) and 3/5 chickens at 9 dpi. (Fig. 4K). No microscopic lesions were observed in the hearts of control chickens (Fig. 4L). No obvious microscopic lesions of brains were observed in the infected chickens ( Fig. 4M-O) or control group (Fig. 4P).

Microscopic lesions. Microscopic lesions are summarized in
In the spleens all infected chickens showed severe lymphocyte depletion, haemorrhage and necrosis at 3 dpi (Fig. 5a). Also, at 6 dpi, many splenic lymphocyte nuclei became pyknotic, and severe focal necrosis was observed (Fig. 5b). Lesions were less pronounced at 9 dpi, with moderate lymphocyte necrosis and haemorrhage (Fig. 5c). No microscopic lesions of spleens were observed in the control group (Fig. 5d). Lesions were not obvious in the duodenum at 3 dpi (Fig. 5e). Duodenums of four infected chickens all showed detachment of mucosal epithelium at the tip of villi and at 6 dpi and 9 dpi. (Fig. 5f,g). No microscopic lesions of duodenums were observed in the control group (Fig. 5h).
The bursa of all infected chickens showed microscopic lesions at 3 dpi, 6 dpi and 9 dpi. Lymphocyte depletion was observed at 3 dpi (Fig. 5i). Heterophilic granulocytes were increased and infiltrating in the cortex at 6 dpi (Fig. 5j). There was more heterophilic granulocytes infiltrating at 9dpi (Fig. 5k). Microscopic lesions in bursa were not present in the control group (Fig. 5l). In the kidney, tubules showed swelling and granularity of the cytoplasm at 3 dpi in 4/5 animals (Fig. 5m). Moderate alterations were observed in 4/5 chickens at 6 dpi, Figure 3. Body weight changes for chickens infected with NDRV by the s.c. route. The experimental group chickens were infected with 100 µL × 10 5.00 ELD 50 /0.1 mL of NDRV. The control group chickens were infected with 100μL PBS (pH 7.0). Bars show means ± standard deviation (SD). The mean value was statistically significant, determined using by the two-tailed Student's unpaired t-test (*P < 0.05 and **P < 0.01). Table 1. Gross lesions in chickens of two different groups at 3 dpi, 6 dpi and 9 dpi. a Number of chickens exhibiting gross lesions in the sacrificed chickens infected with NDRV. b Severity of gross lessions: −, no gross lesions; +, mild gross lesions; ++, moderate gross lesions; +++, severe gross lesions. www.nature.com/scientificreports/ including inflammatory cell infiltration (Fig. 5n). At 9 dpi granularity of the cytoplasm was mild in 4/5 animals with few inflammatory cells were observed (Fig. 5o). No microscopic lesions of kidney were observed in the control group (Fig. 5p).

Organ
Immunohistochemical examinations. Viral antigens in the tissue sections were stained brown by the immunohistochemical examinations at 6 dpi. In the liver, positive immunohistochemical staining signals were mainly distributed in the liver cytoplasm and nuclei. The nuclei of red blood cells within the hepatic sinus gap were also positively stained (Fig. 6A). In the brain, positive staining signals were widely distributed in the cytoplasm of the neurons, and part of the nucleolus was positively stained (Fig. 6C). In the spleen, positive signals were widely distributed in the red pulp. The concrete localization of positive signals was in the nucleus of the splenocytes (Fig. 6E). In the kidney, strong positive signals were mainly focused on the renal tubular epithelial cell cytoplasm and nuclei. Part of the nuclei was strongly positively stained. The glomerular podocyte cytoplasm also presented positive reactions (Fig. 6G). No obvious positive signals were observed in the control groups ( Fig. 6B,D,F,H). The background staining was observed in the control of liver, brain and kidney, but it was very slight (Fig. 6B,D,H).

Viral loads in different tissues.
Viral loads were detected in the heart, liver, spleen, lung, kidney, brain, intestines and bursae of infected chickens by a SYBR Premix EX Taq assay. It can be seen in Fig. 7 that the viral loads in the liver, spleen, lung, kidney and intestines reached the peak at 3 dpi. In particular, the viral loads in the liver, spleen and lung were much higher than those in the other tissues. The level of viral RNA in different tissues began to decline from 6 dpi, but the viral loads in the liver and spleen always remained at a higher level before 6 dpi, and this result was consistent with the mortality of chickens infected with NDRV. The level of viral RNA in the bursae reached a peak at 6 dpi and maintained a higher expression level. Viral loads of the brain always remained at a lower level. No viral RNA was detected in the control group.
Detection of the serum neutralizing antibody. As seen in Fig. 8, chickens infected with NDRV induced neutralizing antibodies at 7 dpi (SN antibody titre > 10). From 7 dpi, the neutralizing antibody titres were continuously ascending, and the level at 14 dpi reached a much higher level. No positive neutralizing antibody titres were detected in chickens in the control group (SN antibody titre < 10).

Discussion
NDRV infection has become a common disease in the Chinese duck industry. Although its mortality rate has not been high, the disease can cause spleen necrosis and immune suppression, which can develop into a serious secondary infection and growth retardation in ducks. More importantly, NDRV infection is a disease that can spread horizontally, as well as vertically, and the offspring of infected breed ducks can easily transmit the virus and cause disease 17 . Therefore, NDRV infection is much more difficult to prevent and control in practical production. NDRV can infect different types of ducks and cause diseases, but research on the pathogenicity of NDRV to chickens has not been reported until now. This study was focused on the clinical symptoms, pathological changes, viral RNA expression, and serum antibodies of infected chickens to further evaluate the pathogenicity of NDRV to chickens. It was reported that NDRV can infect 10-day-old chicken embryos by allantoic cavity inoculation. Chickens were delayed in hatching and had obvious necroses in the liver and spleen 6,16 . In this study, 3-day-old chickens infected with NDRV subcutaneously exhibited body weight loss, introflexion of claws, performing of splits, death, etc. The most typical gross lesions mainly included swelling, brittleness, and yellowish-white focal necroses in the liver and spleen. These experimental data were consistent with the symptoms and lesions of ducks infected with NDRV 6,16,17 . In previous studies, lymphocyte depletion could be observed in most of the tissues of ducks infected with NDRV 14,17,18 . In this study, similar lesions were also found in different chicken organs. For example, inflammatory cell infiltration was severe in the liver, while lymphocyte depletion was obvious and typical in the spleen and bursa. The spleen and bursa are very important immune organs in poultry, pathological damage to which can lead to the immunosuppression of the body. In particular, the bursa plays an important role in inducing B lymphocytes to differentiate and mature, so lymphocyte depletion could lead to immune dysfunction. www.nature.com/scientificreports/ In addition to the liver, spleen and bursae, other tissues also demonstrated varying degrees of pathological changes, such as lymphocyte infiltration with congestion in the lung, cytoplasmic granularity in the myocardium of the heart with lymphocyte infiltration, cytoplasmic granularity in renal tubules, etc. Most studies of the pathogenicity of NDRV mainly focused on the immune organs of poultry, and other organs were rarely involved 7,12,14,15,18,19 . However, in this study most of the organs of the chickens were observed to develop severe lesions, which was in agreement with the pathogenicity study of NDRV to ducks 17 . It was reported that the lesions of brain were slight in ducks 17 . In this study no lesions were observed in the chicken brains. It also means that tissue tropism of MDRV to brain is very slight and there were differences between chickens and ducks.
In this study, the viral RNA of most tissues reached the highest level at 3 dpi, and the level of viral RNA expression began to decline from 6 dpi. The data suggested that the developing speed of the disease was much faster, consistent with the deaths among diseased chickens being mainly focused in the period of before 6 dpi. The results were also consistent with the dynamic changes in viral loads in duck tissues 17 . In this study, pathological changes in the spleen and liver were much more serious at 6 dpi, and they were closely related to the level of the viral loads being maintained for a longer time in these tissues. The viral loads of the liver and spleen were much higher than those in other tissues and were maintained for a longer period, consistent with the severe and obvious lesions in the two organs. Compared to other tissues, the viral loads of the bursae reached the peak level at 6 dpi and always maintained a higher level until 9 dpi. These data indicated that the liver, spleen and bursae of chickens could be the target organs of NDRV.
The viral loads of the brain always maintained at a lower level which consistent with the pathological characteristics of the brain. www.nature.com/scientificreports/ In this study, the chickens could produce neutralizing antibodies to NDRV at 7 dpi. This finding indicated that NDRV could infect chickens subcutaneously and induce neutralizing antibodies.
This study indicates that chickens can be infected subcutaneously with a virulent NDRV strain that can cause disease or even death. However in practical poultry production chickens and ducks were rarely mixed, so chickens were not easy to be exposed to the virus. Even if chickens were exposed to NDRV, it was very difficult for chickens to infect a high-dose virus a time in practical poultry production. Therefore, there have been no cases infected with NDRV in chicken production. But if the frequent exposure to NDRV always remains, which could lead to virus variation and the virus will be probably susceptible to chickens. And that reovirus is a multi-segment RNA virus 1 , chicken-derived reovirus and duck-derived reovirus are more likely to exchange and recombine. It suggests that the virus may infect chickens naturally in practical poultry production in the future. Therefore in the actual production, chickens and ducks should avoid being mixed to breed and thus chickens will avoid being exposed to the duck-derived viruses, such as NDRV. And chickens will also avoid adapting to the virus due to long-time exposure of NDRV. Although so far chickens have not been naturally infected with the NDRV strain, there is still a great risk of infection. This study provided some experimental data for further prevention and control of infection by NDRV in poultry.

Materials and methods
Animals and virus. Three-day-old SPF chickens and 9-day-old SPF duck embryos were purchased from Shandong Hao Tai Experimental Animal Breeding Company Limited. The novel duck reovirus (NDRV) strain was isolated from one Cherry Valley duck farm in Shandong Province in 2012. The virus was named SD-12 (Accession number: KJ879930). It was inoculated into the allantoic cavities of 9-day-old SPF duck embryos for subculture. After three passages, the virus was used for the challenge in this study. The virus titre was 10 5.00 ELD 50 /0.1 mL (50% lethal dose for embryos) for the stock, determined according to the method of Reed and Muench 20 . The virus was isolated by our team members. PCR detection showed that there were not duck enteritis virus, duck hepatitis A virus, duck tembusu virus, duck parvovirus and other common duck virus in the allantoic fluid.
Animal experiments. Eighty 3-day-old SPF chickens were raised in negative pressure isolators and randomly divided into two groups, with 40 chickens in each group. The experimental group was inoculated with 100 www.nature.com/scientificreports/ μL of allantoic fluid (10 5.00 ELD 50 /0.1 mL) of NDRV by the s.c. route. The other group, as the controls, was inoculated with 100 μL of PBS. The virus dose was determined by a pre-test. Clinical symptoms of the two groups were observed every day. Ten chickens were randomly selected from each group and weighed every three days for 14 days. Blood of the infected chickens was collected and analysed every day. At 3, 6, 9 dpi, five chickens from each group were euthanized by carbon dioxide, and their tissues (heart, liver, spleen, lung, kidneys, brain and intestine) were collected. One part of the tissue sample was fixed in 10% neutral buffered formalin solution for histological examination. The other part of the tissue sample was stored at − 80 °C until use for RNA extraction. The test animal bodies, embryo bodies and the used test materials were disposed harmlessly. The whole animal experiments were conducted in the Biosafety Level 2 laboratory.  . Serum neutralizing antibody against NDRV was detected. Each sample was tested in triplicate. Serum neutralizing antibody titers were expressed as the reciprocal of the log2 of the highest dilution serum dilution that inhibited 50% CEF death. The negative antibody titer was zero. Data are expressed as mean ± SD (n = 3). Significant differences were determined by the two-tailed Student' s unpaired t test. Histopathology and immunohistochemistry examinations. Tissue samples were fixed in 10% neutral buffered formalin solution for 72 h, dehydrated, and embedded in paraffin wax. Sections that were 4 µm thick were cut. One part of the sections was stained with haematoxylin and eosin (H&E) following the standard histopathological protocols, and the pathologic results were observed under a microscope. The other part of the sections at 6 dpi was deparaffinized with xylene and hydrated with different grades of alcohol liquid (100-75%) for the immunohistochemical examination. 0.01 M sodium citrate buffer solution (pH6.0) was heated to 95 °C and then sections were immersed for 10 ~ 15 min to retrieve the antigen. After blocking with 5% goat serum albumin buffers for 1 h, the sections were incubated with rabbit sera against NDRV overnight at 4 °C. After three washes with PBS, the sections were conjugated with a diluted mouse anti-rabbit HRP-conjugated polyclonal serum for 1 h at 37 °C. Diaminobenzidine could be used as the substrate chromogen. After counterstaining with haematoxylin, the sections were sealed with neutral gum and observed with the microscope. Rabbit sera against NDRV were prepared by our own laboratory. Purified and concentrated NDRV was used as the immunological antigen. The acquired antisera on rabbit anti-NDRV didn't react with the main chicken viruses such as newcastle disease virus, avian influenza virus, infectious bronchitis virus, infectious laryngotracheitis virus, infectious bursal disease virus, avian leukosis virus and chicken reovirus by a serum neutralization test. The antisera also didn't react with the main duck viruses such as duck enteritis virus, duck hepatitis A virus, duck tembusu virus, duck parvovirus and other common duck virus by a serum neutralization test. It was diluted by 1:200 and used as the primary antibody of the immunohistochemistry examinations.
Analysis of serum antibodies against NDRV. Serum neutralization testing (SNT) was used to detect the serum antibody titres against NDRV. Serum samples of three chickens were randomly collected respectively at 0 dpi, 7 dpi and 14 dpi, respectively. Before starting the tests, complements in serum samples had to be inactivated at 56℃ for 30 min. SNT was performed with duck embryo fibroblasts (DEF) as previously described 12,17 . Serum neutralizing antibody titres were expressed as the reciprocal of the log2 of the highest serum dilution that inhibited 50% DEF death and was calculated by the method of Reed and Muench 20 . Each sample was performed in triplicate.
Statistical analysis. The experimental test data are expressed as the means ± standard deviations. Serum antibody titres and body weight data were analysed with Student's two-tailed unpaired t-test. Viral loads in different tissues were evaluated by one-way analysis of variance (ANOVA) with Tukey's post-test. Statistical significance was represented by P < 0.05 and P < 0.01.